Human-Centered Artificial Intelligence with Applications to Smart Manufacturing and Healthcare Systems

Tremendous artificial intelligence (AI) methods have been invented aiming to reduce workforce demands in labor-intensive tasks, such as manufacturing process monitoring and visual inspection for process and quality control. Although the goal of smart manufacturing is to achieve fully automated data-driven decision-making, existing AI methods are far from sufficiently accurate and flexible to replace human workers. An intuitive way to improve AI’s performance is to “incubate” AI by showing how an experienced human worker makes a decision step by step. However, the natural gap between the complex AI model structures and human domain knowledge structures prevents this incubation. The gap can be attributed to the lack of a fundamental human-AI mutually interpretable method to share and preserve knowledge between human and AI. In this talk, I present an interpretable neural network (INN) method that bridges high performance AI models with humans’ domain knowledge in quality modeling of a semiconductor manufacturing process and in perioperative blood pressure management. I will also cover a summary of my other research studies in human-centered AI.

Dr. Xiaoyu Chen is an Assistant Professor of Industrial and Systems Engineering at University at Buffalo. He received his Ph.D. degree from the Grado Department of Industrial and Systems Engineering and earned his M.Eng. degree from Computer Science at Virginia Tech in 2021. Dr. Chen is a member of Institute for Operations Research and the Management Sciences (INFORMS), Institute of Industrial and Systems Engineering (IISE), and Association for Computing Machinery (ACM). His ongoing research projects are funded by industry partners, state government agency, and American Heart Association.